1137801764 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net 1137802143 J * hillct ~H@client200-5.dsl.intrex.net 1137802189 M * hillct evening all 1137802269 M * daniel_hozac evening 1137802284 M * hillct hi 1137802289 M * hillct how goes it? 1137802313 M * daniel_hozac fine, you? 1137802324 A * hillct really should add ScoopHost to the list of vendors using linux-vservers 1137802332 M * hillct I have a stupid problem 1137802339 M * hillct that's more bash than vserver 1137802352 M * hillct but if you don't mind... 1137802375 M * hillct I'm trying to exec a command within a vserver which contains output redirection 1137802404 M * daniel_hozac output redirection to the host or the guest? 1137802413 M * hillct obviously this doesn't work 1137802414 M * hillct vserver infosecexchange.com exec echo 'source /etc/profile.d/prompt.sh' >>/etc/skel/.bashrc 1137802425 M * hillct because stdout is redirected on the host 1137802429 M * daniel_hozac right. 1137802432 M * hillct I want to redirect it within the guest 1137802445 M * hillct qraping it in quotes doesn't do it 1137802452 M * daniel_hozac vserver ... exec bash -c '... > ...' should do it. 1137802458 M * hillct ah 1137802472 M * hillct hmm 1137802496 A * hillct tries that 1137802597 M * hillct works like a champ. thanks 1137804291 P * hillct Kopete 0.11 : http://kopete.kde.org 1137804517 N * Bertl_oO Bertl 1137804522 M * Bertl back now ... 1137804805 J * waldi ~waldi@bblank.thinkmo.de 1137804889 M * kilian hi guys 1137804902 M * kilian Bertl: do you reckon there is no 2.6.16rc1 patch out yet? 1137804986 M * kilian Bertl: i.e. latest linux version is 2.6.15? 1137805031 M * Bertl well, there is a 2.6.16-rc* on kernel org, but you mean for linux-vserver, right? 1137805043 M * kilian of course =) 1137805056 M * Bertl haven't bothered yet ... 1137805058 M * kilian i mean "latest linux version supported by linux-vserver" 1137805067 M * kilian uhm, how come? 1137805113 M * Bertl the early -rc's are usually not stable enough to test anything 1137805124 M * Bertl and currently I'm trying to test a few things ... 1137805146 M * Bertl and, there was no demand yet ... 1137805265 M * kilian mmh, would i count as "demand"? ;) 1137805280 M * Bertl sure, you would like to have one? 1137805288 M * kilian Yes. =) 1137805383 M * Bertl hmm, okay ... will take a little as I have to get my backup done first, but should be there in a few hours ... 1137805395 M * kilian great. thnx! 1137805402 M * Bertl kilian: any feedback regarding dietlibc on hppa? 1137805411 M * kilian Bertl: none that i have heard of 1137805419 M * Bertl nothing here either ... 1137805935 M * jpacheco hey everyone :) 1137805998 M * Bertl hey! 1137806044 M * jpacheco anyone here watch starget atlantis? 1137806057 M * _Roey Berteleh hi 1137806058 M * jpacheco cause im gonna sit back and watch ep18 :) 1137806067 M * jpacheco stargate* 1137806273 M * Bertl well, ahve fun then! 1137806331 P * undefined 1137808288 M * Bertl kilian: btw, who is the debian maintainer of dietlibc? 1137808403 M * DaCa Bertl: Maintainer: Gerrit Pape 1137808416 M * Bertl ah, tx ... 1137808553 M * DaCa in general, if you dont have debian installed, you can find this kind of info on http://packages.debian.org/packagename 1137808632 M * Bertl ah, okay, good to know ... 1137808662 M * Bertl let's see if I can remember that next time :) 1137808689 M * DaCa thats what your browser has bookmarks for :) 1137808726 M * Bertl hmm, no, the boormarks are for important stuff :) 1137808758 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net 1137808857 M * Bertl hmm, ebiederm will have fun forward porting his patches 1137808873 M * daniel_hozac lots of changes? 1137808888 M * Bertl 2.6.14 to 2.6.15 was a huge pile of changes 1137808901 M * Bertl seems that 2.6.15 to 2.6.16 is quite huge too 1137808909 M * daniel_hozac heh, and that's just rc1... 1137808929 M * Bertl well, rc1 means that the first settlement stage has begun 1137808949 M * Bertl but yeah, the 2.6.16 is likely to change a lot more 1137809807 M * Bertl well, seems they incorporated parts of BME, of course in several 'trivial' patches :) 1137809990 M * daniel_hozac but not all of BME? 1137810006 M * Bertl the noatime and nodiratime 1137810051 M * daniel_hozac ah. 1137815730 Q * lilalinux Remote host closed the connection 1137816021 Q * mkhl Quit: 1137816052 Q * pusling_ Read error: Connection reset by peer 1137816127 J * pusling pusling@195.215.29.124 1137818814 P * undefined 1137822617 N * nokoyaz nokoya 1137822659 M * Bertl morning nokoya! 1137826300 M * Hollow morning 1137826471 M * _Roey Bertl: where are you 1137826475 M * _Roey Hollow: where are you, too? 1137826478 M * _Roey germany? 1137826693 M * Hollow yep 1137826710 M * Bertl _Roey: austria 1137828138 J * Doener doener@i5387C8EC.versanet.de 1137828147 M * Doener morning 1137828151 M * Bertl morning! 1137828375 M * Doener having not really followed channel traffic lately, did ebiederm join development? just noticed yesterday(?) that you were talkink about ipv6 stuff? 1137828429 M * Bertl well, ebiederm is working for a company, and his main goal is to get virtualization stuff into mainline, similar to Serge (Hallyn) 1137828485 M * Bertl nevertheless, I think we could make good use of some of the 'namespaces' he plans to add, given that they are general enough while still efficient 1137828504 M * Bertl that is why we are spending a lot of time on discussing the details 1137828505 M * Doener serge was the container/freeze guy (ibm?)? (trying to sort information ;) 1137828524 M * Bertl Serge = linux-jails 1137828681 M * Doener hm, but it's the same that posted the pid virtualization patch to lkml, right? 1137828873 M * _Roey Bertl: ahhhhhhh ok 1137828875 M * _Roey Hollow: cool. 1137828876 M * _Roey toll. 1137828893 M * Hollow ?? 1137828983 M * Bertl Doener: well, yes ... 1137829015 M * _Roey Hollow: toll = cool auf Deutsch 1137829020 M * _Roey unless they changed it 1137829020 M * _Roey ? 1137829061 M * Hollow hm, yeah, but what is cool? did i miss something? 1137829163 M * _Roey just saying 'ok' 1137831373 J * kjo nobody@news.stud.uni-goettingen.de 1137831571 Q * kjo Quit: 1137833131 M * Bertl okay, another round for BME ... :) 1137833156 M * Bertl off to bed now ... back later in the evening ... 1137833180 N * Bertl Bertl_zZ 1137834169 J * Viper0482 ~Viper0482@p549761A9.dip.t-dialin.net 1137834258 Q * yang Ping timeout: 480 seconds 1137834571 M * _Roey gnight 1137836405 J * meandtheshell ~markus@85-124-15-55.dynamic.xdsl-line.inode.at 1137839364 Q * dlippolt Ping timeout: 480 seconds 1137840296 J * bonbons ~bonbons@83.222.39.249 1137840932 J * Smutje ~Smutje@xdsl-84-44-241-216.netcologne.de 1137841039 Q * Smutje_ Ping timeout: 480 seconds 1137842808 M * lonewolff morning all 1137842816 M * daniel_hozac morning 1137845212 J * menomc ~amery@200.75.27.85 1137845321 Q * mnemoc Ping timeout: 480 seconds 1137845321 N * menomc mnemoc 1137847584 M * Hollow bonbons: ping 1137847639 M * bonbons Hollow: pong 1137847677 M * bonbons Hollow: please ping again in a good hour (lunch time here) 1137847711 M * Hollow ok 1137848226 Q * pusling Read error: Connection reset by peer 1137848228 J * pusling_ pusling@195.215.29.124 1137848435 Q * lonewolff Read error: Connection reset by peer 1137848470 J * lonewolff ~lonewolff@host-84-9-143-159.bulldogdsl.com 1137848517 Q * mountie Remote host closed the connection 1137848999 J * undefined ~undefined@adsl-68-93-109-94.dsl.rcsntx.swbell.net 1137851832 M * bonbons Hollow: ping (I'm back) 1137852434 J * mountie ~mountie@CPEdeaddeaddead-CM000a739acaa4.cpe.net.cable.rogers.com 1137853605 Q * lonewolff Read error: Connection reset by peer 1137853685 J * lonewolff ~lonewolff@host-84-9-143-159.bulldogdsl.com 1137855097 M * Hollow bonbons: back too ;) 1137855114 M * bonbons Ok :) 1137855145 M * Hollow ok, so i you have some time we could try to debug the failed to create network context error.. 1137855161 M * Hollow though i'm a bit helpless until now 1137855163 M * Hollow but well.. 1137855197 M * Doener debug? sounds like fun :) (unless too much bash is involved ;) 1137855205 M * bonbons I have the time, just need to start the box and re-replace util-vserver by libvserver+vserver-utils 1137855230 M * Hollow Doener: yeah, sounds like real fun, because libvserver seems to use the incorrect syscall matric number.. 1137855302 M * Doener Hollow: is there any information available, yet? 1137855325 M * Doener (like strace etc.) 1137855356 M * bonbons Hollow: should I build the lib manually / use the most up-to-date version from svn? 1137855365 M * Hollow hm, it's been a while since the error was reported, dunno 1137855398 M * Hollow bonbons: 1.0.1 is the same as svn trunk 1137855399 M * bonbons Doener: last week-end I provided a strace, but don't know if any additionnal info came out since then 1137855446 M * Doener http://homepage.internet.lu/brunop/strace_vncontext this one? 1137855484 M * Hollow yup 1137855504 M * Hollow 0xb7f1a3ac seems somewhat fucked up, no? 1137855543 M * Doener that'd be nx_Create in libvserver? 1137855553 M * Hollow yup 1137855652 M * bonbons Hollow: does libvserver have some flag to enable extra debugging code? 1137855734 M * Hollow CFLAGS="-g -ggdb3" ./configure 1137855735 M * Hollow ;) 1137855818 M * bonbons configure script says 'syscall number 273' 1137855852 M * Hollow yeah, the syscall is vserver() and in the trace, but the first argument is the correspoding vserver command 1137855857 M * Hollow VCMD_net_create in this case 1137855963 M * Hollow bonbons: do you compile manually or via portage? 1137856014 M * bonbons I did compile via portage, but am now looking at it for manual compilation (to replace what portage installed) 1137856148 M * bonbons By the way, the tons of output by vserver-utils's init script is buggy as well? (always Function not implemented messages) 1137856197 M * Hollow bonbons: hm... sounds like other syscalls are failing too? because the init script does not call net_create afaik 1137856264 M * bonbons probably they fail all together... 1137856286 M * Hollow very weird 1137856315 M * Hollow so, do you have the libvserver source tree at hand? 1137856399 M * bonbons yup, and was grepping around after configure to find out where the sys-call number is used, but no success 1137856416 M * Hollow ok, did you specify the CFLAGS above? 1137856452 M * bonbons config.h defines __NR_vserver, but it's never used. Yup, with CFLAGS from above 1137856490 M * Hollow ok, let me take a look where it should be used ;) 1137856554 M * Hollow well, it is used in syscall.h if you grep for _NR you see it is concatenated from the second argument to the syscall macro which is in syscall.c 1137856624 M * bonbons The __sr_id(N) define? 1137856639 M * Hollow yep.. but the syscall number (273) is not the problem here, since strace verifies the vserver syscall was called 1137856750 M * Doener bonbons: is that on x86? 1137856784 M * Hollow yup 1137856813 M * bonbons Yes, x86, on Pentium M => arch=i686, march=pentium-m 1137856820 M * Hollow any strange optimizations? 1137856846 J * NikDaPhreak ~NikDaPhre@193.24.241.34 1137856858 M * NikDaPhreak hi all 1137856944 M * bonbons I just replaced the lib with a manually compiled one and there it works! 1137856976 M * bonbons Hollow: I guess the strage optimizations will be the right path for searching... 1137856990 M * Hollow damn ricers :P 1137857028 M * Hollow but it would be nice to know which flag it is, so i can filter it in the ebuild 1137857044 M * Hollow could you paste the output of "emerge info" somewhere? 1137857150 M * bonbons http://homepage.internet.lu/brunop/emerge.info 1137857166 M * Hollow uhm, well 1137857173 M * Hollow no ricer flags 1137857192 N * Bertl_zZ Bertl 1137857196 M * Bertl morning folks! 1137857200 M * Hollow morning bertl 1137857212 M * Hollow Bertl: right in time ;) 1137857226 M * Bertl am i? 1137857230 M * bonbons morning Bertl 1137857240 M * Hollow Bertl: we're trying to debug strange syscall issues 1137857284 M * Hollow http://homepage.internet.lu/brunop/strace_vncontext 1137857390 M * Bertl hmm, that's x86, right? 1137857394 M * Hollow yep 1137857419 M * Bertl could you upload the binary or the .S file of your compile? 1137857432 M * bonbons with CFLAGS and compiler settings as in http://homepage.internet.lu/brunop/emerge.info 1137857445 M * Bertl hey bonbons! 1137857477 M * Hollow the bin would be libvserver.so ? 1137857531 M * bonbons Bertl: what CFLAGS to you want exactly? PS: I will first try if it's one that spits out not implemented function or not 1137857533 M * Bertl hrm, no, in this case both, the executable and the lib, but, does it still happen when you compile it static? 1137857551 M * Hollow Bertl: well, he compiled oit manually, and it worked 1137857561 M * Hollow so we concluded the CFLAGS are the evil thing 1137857581 M * Bertl well, I'd look into the omit-frame-pointer then 1137857602 M * Bertl (but nevertheless it looks to me like something in the code is wrong) 1137857630 M * Hollow ok, guess you have to debug it with bonbons since i cannot reproduce it 1137857648 M * Hollow though... if i compile it omit-fp 1137857651 M * Hollow i'll try 1137857693 M * Bertl okay, as I said, will require either the binar(y/ies) involved or the assembler output of the relevant sscall 1137857702 M * Bertl *syscall 1137857704 M * Hollow ok, just a sec 1137857738 M * bonbons Looks like without -fomit-frame-pointer it does not fail 1137857809 M * bonbons Just manually recompiled once with, once without -fomit-frame-pointer and with it, Function not implemented errors occur (with [un]hide procfs) 1137857895 M * Hollow Bertl: http://home.xnull.de/misc/libvserver.objdump http://home.xnull.de/misc/vcontext.objdump 1137857904 M * bonbons bertl, which files to you want? libvserver.a, libvserver.so, also one of the binaries (they do not influence as I was just replacing libvserver.so) 1137858139 M * Bertl Hollow: shouldn't that be vncontext or does vcontext fail too? 1137858147 M * Hollow every syscall fails 1137858157 M * Hollow (i did not try all, but some) 1137858195 M * Bertl okay, good :) 1137858206 M * Hollow ;) 1137858392 M * Bertl was the library compiled with omit-frame-pointer too? 1137858398 M * bonbons Note, mainting the -O2 option seems to work, though from gcc's man page it should include -fomit-frame-pointer ... 1137858404 M * Hollow only the library was 1137858517 M * Doener bonbons: IIRC -O2 includes -fomit-frame-pointer on x86-64 only 1137858578 M * Doener -O also turns on -fomit-frame-pointer on machines where doing so 1137858578 M * Doener does not interfere with debugging. 1137858589 M * bonbons Doener: ok, so the man-page is a bit too incomplete => incorrect :) 1137858599 M * Doener the -fomit-frame-pointer descriptions lacks that information 1137858808 M * bonbons Bertl: I did generate two objdumps of libvserver, once with -fomit-fp once without, so comparison should be easy 1137858859 M * Bertl bonbons: I think you cannot mix code with omit-frame-pointer with non omit-frame-pointer code 1137858930 M * Bertl at least not without saving the ebp between calls 1137859000 M * bonbons you mean app with omit-fp can't work with lib without? or just inside a same binary? 1137859030 M * Bertl yes, that's what I think ... 1137859109 M * bonbons hmm... would have to check how vserver-utils get built then (not that the make-file of one of both strips out omit-fp and not the otherone) 1137859176 M * bonbons Anyhow, my two objdumps of libvserver can be found here: http://homepage.internet.lu/brunop/libvserver.so-omitfp.objdump, http://homepage.internet.lu/brunop/libvserver.so.objdump 1137859195 J * lilalinux ~plasma@80.69.35.186 1137859235 Q * shedi Quit: Leaving 1137859258 M * Hollow bonbons: but if you install both libvserver and vserver-utils both should be compiled with omit-fp, no? 1137859452 M * bonbons thet's what I guess, both were initially installed by protage (should be with omit-fp) and don't work. When I replaced libvserver with the one manually compiled without omit-fp it does not spit out function not implemented. 1137859503 M * bonbons So unless one of the Makefiles/Configure scripts filters out CFLAGS that the otherone does not filter out, they have the same flags 1137859523 M * Bertl well, both versions look perfectly fine to me ... 1137859574 M * Bertl but I do not expect an ebp binary to work with the ofp library 1137859596 M * bonbons Bertl, and the other way round? 1137859620 M * Bertl other way round should be fine 1137859640 M * Hollow and if both are ofp? 1137859678 M * Bertl should be fine too, but I'm currently wondering if that isn't another new gcc feature 1137859696 M * bonbons I'm now compiling vserver-utils manually, once with omit-fp, once without 1137859696 M * Hollow well, if he installed it via portage, both have ofp, and it didn't work.. 1137859736 M * Bertl Hollow: which version of the 'shiny' does your code use? 1137859751 M * Hollow good question.. 1137859753 M * Bertl i.e. could you point me to the relevant file? 1137859757 M * Hollow sure 1137859775 M * Hollow http://dev.croup.de/proj/libvserver/browser/trunk/src/syscall.h 1137859828 M * Bertl bonbons: could you dump the defined symbols with and without ofp for me? 1137859920 M * bonbons Bertl, symbols of which binary/lib? I'm not yet expert in dumping :), but have vncontext and libvserver in both version now 1137860038 M * Bertl add -E -dM to the gcc with any file and one time -fomit... the other time without ... 1137860063 M * Bertl should dump you a bunch of defines 1137860100 M * bonbons You mean at compile-time, or on already sxisting binaries? http://homepage.internet.lu/brunop/vserver.tar with both versions of vncontext and libvserver.so 1137860128 M * Bertl compile time, e.g. gcc -O2 -E -dM test.c 1137860230 M * Bertl gcc -O2 -E -dM - : with -ofp 1137860842 M * Bertl (from the library) 1137860885 M * bonbons doing it... 1137860969 M * bonbons make fails... 1137861002 M * Bertl with? 1137861035 M * bonbons ups, did copy edp insteqd of ebp ... 1137861242 M * bonbons no copilation works with -fomit-fp, but not without! syscall.c:40:error: bp cannot be used in asm here 1137861333 M * Bertl well, that's the problem, but it seems your gcc does not tell when it uses the -fofp 1137861357 M * bonbons Objdump with omitfp is here: http://homepage.internet.lu/brunop/libvserver.so-omitfp2.objdump 1137861358 M * Bertl so it's not possible to work around that, without adding a special 'define' case 1137861446 M * Bertl yep, I'd assume the 'new' version should work quite fine 1137861447 M * bonbons but the change doesn't fix anything 1137861451 M * Bertl no? 1137861493 M * bonbons no, I get the function-not-implemented results with it 1137861518 M * Hollow yep.. can confirm this 1137861642 M * Bertl what about the libc? 1137861668 M * Hollow 2.3.6 here 1137861671 M * Bertl does it work around the missing ebp? 1137861707 M * Hollow hm, how is libc involved here? 1137861707 M * Bertl I would prefer to test the following case: 1137861729 M * Bertl library .. compiled with -fofp 1137861739 M * Bertl binary .. compiled without -fofp 1137861745 M * Bertl glibc unknown 1137861764 M * Bertl check this with the old version of syscall.h 1137861789 M * Bertl if that doesn't fail, then the problem seems to reside in your libc 1137861855 M * Hollow Bertl: fails 1137861883 M * Bertl the kernel syscall uses %ebp for the sixt argument, but userspace uses %ebp for the framepointer 1137861905 M * Bertl okay, if you use the new syscall.h version, does it work then? 1137861935 M * Hollow still fails 1137861951 M * Bertl hmm .. strange, but points into a different direction then 1137861960 M * Bertl I would first get rid of the libc 1137862144 Q * pusling_ Read error: Connection reset by peer 1137862146 J * pusling pusling@195.215.29.124 1137862177 M * Hollow getting rid of libc means? 1137862201 M * bonbons I have lib 2.3.5 1137862214 M * Bertl use dietlibc if possible, or compile it static 1137862238 M * Bertl I would have to dig through the entire glibc you have to find the code 1137862247 M * bonbons an dietlibc 0.28 is installed too (would need to check which one it links against) 1137862267 M * Bertl dietlibc links at compile time 1137862284 M * Bertl (i.e. produces static binaries by default) 1137862356 M * bonbons here they are linked against glibc 1137862373 M * Bertl yes, I saw that in the obj-dump :) 1137862533 M * Hollow ../../src/syscall.c: In function 'sys_vserver': 1137862533 M * Hollow ../../src/syscall.c:40: error: bp cannot be used in asm here 1137862538 M * Hollow compiling with diet 1137862562 M * Bertl without omit-frame-pointer I assume? 1137862569 M * Hollow yes 1137862582 M * Bertl okay, and with? 1137862608 M * Bertl problem is, the -fofp seems not detectable 1137862622 M * Hollow compile works 1137862631 M * Bertl i.e. you would have to pass something like -DOFP or so :) 1137862636 M * bonbons Hollow: same as what I got here... looking where I need to operate to use dietlibc 1137862686 M * Bertl Hollow: now the interesting question is, does the dietlibc vcontext fail too (with -fofp) 1137862713 M * Hollow first question is why configure fails in vserver-utils if i use diet ;) 1137862730 A * Bertl .o( not that I think -fofp is such a good idea ) 1137862748 M * Hollow well, i can live with just filtering it ;) 1137862792 M * Bertl probably that's the best for now anyway 1137862807 M * Bertl (as long as we do not know how to figure the option from gcc) 1137862836 M * bonbons Hollow: what do I need to to to make libvserver & vserver-utils use dietlib? 1137862882 M * Hollow CC="diet gcc" ./configure ... 1137862899 M * Hollow also use --disable-shared with libvserver 1137862931 M * bonbons ok 1137862990 J * vrwttnmtu ~eryktyktu@82-69-161-137.dsl.in-addr.zen.co.uk 1137862991 M * Hollow hm, seems like dietlibc doesn't have forkpty 1137863002 M * Bertl welcome vrwttnmtu! 1137863012 M * vrwttnmtu :) 1137863013 M * Bertl Hollow: it's in libcruft or so? 1137863046 M * Hollow hm, there is openpty 1137863051 M * Hollow and i don't have libcruft here 1137863068 M * Bertl sec 1137863139 M * Bertl yep, seems forkpty() is missing 1137863150 A * Hollow sighs 1137863186 M * Bertl it's not part of libc it seems, it's in -lutil, no? 1137863191 M * Hollow yep 1137863209 M * bonbons libvserver and vserver-utils compile fine with dietlibc ... both compiled the same way 1137863247 M * Hollow yeah, fails here because of the new vlogin.c 1137863280 M * Hollow will make it diet compatible 1137863283 M * Bertl why doesn't linking with -lutil work with diet? 1137863287 M * bonbons And using dietlibc it works with omit-fp (still patched syscall.h) 1137863311 M * Bertl okay, so IMHO the glibc is not prepared for -fofp 1137863464 M * bonbons rebuilding without patch still using omitfp works as well 1137864658 J * dlippolt ~dlippolt@cpe-70-112-77-129.austin.res.rr.com 1137864827 M * Bertl welcome dlippolt! 1137865408 M * bonbons Bertl: is it logical that "only" libvserver fails if built with omitfp when glibc is built without, but other apps like mount don't? (probably because of glibc-wrapped syscalls) 1137865507 M * Bertl well, good question, I have no answer atm 1137865577 M * bonbons I'm currently rebuilding glibc with -fomit-frame-pointer and will rebuild libvserver and vserver-utils afterwards to check if it helps. 1137865616 M * bonbons But what I don't understand yet is why util-vserver does NOT fail (does it strip omitfp?) 1137865715 M * Bertl absolutely no idea ... I'm not even sure we are looking at the 'correct' issue yet .. what we know is: 1137865731 M * Bertl - -fofp makes %ebp available for code 1137865749 Q * pusling Remote host closed the connection 1137865753 M * Bertl - -fofp isn't detectable at compile time 1137865790 M * Bertl - a messed up %ebp will happen if you do a syscall without saving %ebp in -fofp case 1137865795 J * pusling pusling@195.215.29.124 1137865904 M * bonbons Let's see what happens once I have glibc compiled with omitfp... 1137866505 M * bonbons Looks like portage strips away omitfp from CFLAGS for glibc 1137866672 M * Bertl maybe libvserver should do so too then ... 1137866690 J * yang ~boni@cpe-212-18-59-124.dynamic.amis.net 1137866745 M * Hollow Bertl: yup, already in my working copy, but i'm adding some diet comfort too atm 1137866751 M * bonbons still wondering how util-vserver does it to work fine 1137866768 M * Hollow it uses dietlibc haha 1137866781 M * bonbons or does it only link to dietlibc? (which works fine) 1137866785 M * Hollow yep 1137866897 M * Bertl yes, I would definitely prefer if the tools would stay dietlibc friendly ... as it eases life for testing and porting 1137867025 M * bonbons So vserver-utils will need to get dietlibc-support in configure + appropriate updates to ebuild (Gentoo) => filter out omitfp unless compiling with dietlibc 1137867069 M * Hollow no, it will get this update to svn :P 1137867076 M * Hollow and a new 1.0.2 will get a new ebuild 1137867163 M * bonbons I mean ebuild for gentoo, configure for svn 1137867185 M * bonbons in the mean-time I'm patching the ebuild locally 1137868606 M * bonbons Hollow: the vserver-utils ebuild (and make install) forgets to create the requirted dir /var/lock/vservers/ 1137868961 M * Bertl okay, off for now ... back later 1137868969 M * Bertl kilian: tests are still running ... 1137868976 N * Bertl Bertl_oO 1137869006 M * ebiederm I am wondering if using the syscall function works better than building an assembly language stub for your weird library compile case? 1137869026 Q * Doener Ping timeout: 480 seconds 1137869038 N * Bertl_oO Bertl 1137869060 M * Bertl ebiederm: the problem with the 'syscall' function is that it is not implemented properly in many archs 1137869062 J * Doener doener@i5387CBE8.versanet.de 1137869064 M * kilian Bertl: thanks for the update ;) 1137869093 M * Bertl ebiederm: so, depending on what *libc* version you are facing, it might screw up things quite nicely 1137869098 M * ebiederm Bertl Ok. I've been using it in /sbin/kexec and no one has complained yet. 1137869117 M * Bertl ebiederm: on what archs? 1137869152 M * ebiederm Mostly powerpc and x86 32 and 64bit variants. It's not enough to be a good sample. 1137869166 M * Bertl okay, that explains it :) 1137869200 M * Bertl really off now ... but would appreciate a chat later ... 1137869205 N * Bertl Bertl_oO 1137870546 M * Hollow bonbons: yeah, will fix that too 1137871297 J * shedi ~siggi@inferno.lhi.is 1137871478 M * bonbons Hollow: I'm trying to get one of my guests running with vserver-utils, but it's not easy... 1137871500 M * Hollow which problems do you have? 1137871594 M * bonbons when starting the guest there was the issue about lock dir, now when trying to stop it, I always get fs.lock: lockfile ... exists; entering namespace fails (I was missing appropriate IP address an startup as well) 1137871645 M * Hollow yeah, the lockfile thing needs to be rewritten 1137871650 M * Hollow it's on my todo list 1137871722 M * bonbons how to I terminate the guest manually? In addition, vps script does not complain on unknown init style 1137871837 M * Hollow vkill -x 1137871909 M * bonbons looks like the guest was only partially created (missing namespace, probably also network context), /proc/virtual/xid/status says it has 1 process, vkill -x xid says no such process 1137871933 M * jpacheco hey everyone 1137871960 M * Hollow bonbons: ok, one second, i'm just committing the dietlibc autoconf thingies and some cleanups and will look at it then 1137871971 M * bonbons ok 1137871992 M * jpacheco what's this dietlibc i've been hearing so much about? 1137872008 M * Hollow www.fefe.de/dietlibc 1137872014 M * Hollow hm 1137872016 M * Hollow 4040 :) 1137872018 M * Hollow 404 1137872036 M * bonbons jpacheco: dietlibc is a minimalistic libc, much smaller than glibc and used e.g. for busybox 1137872050 M * Hollow lol.. dump server 1137872054 M * Hollow no dir redirection 1137872061 M * Hollow *dumb 1137872133 M * Hollow bonbons: ok, fixed the init style error 1137872214 M * Hollow {libvserver,vserver-utils} svn trunk now also has automagic dietlibc detection 1137872510 M * bonbons Somehow I got rid of my "zombie" context... (don't know if it's some timeout or what made it die) 1137872604 Q * yang Ping timeout: 480 seconds 1137872656 M * bonbons now when executing "vserver start myguest" I get "Failed to create networkcontext: No such process" 1137872670 M * Hollow bonbons: kernel/vserver version? 1137872695 M * bonbons 2.6.15-vs2.1.0 1137872731 M * Hollow you need at least 2.1.0.4 1137872746 M * Hollow because of persistant network contexts 1137872790 M * jpacheco bonbons: busybox? 1137872810 M * bonbons Hollow: ok, will upgrade the kernel and try again afterwards 1137872860 M * bonbons jpacheco: busybox is a collection of utilities like cp, mv, mount, ... which fist of rescue-disks (e.g. floppy) 1137872920 M * jpacheco gotcha 1137872929 M * bonbons jpacheco: see http://www.busybox.net/ 1137872936 M * jpacheco i thought you were using it in your vserver 1137872950 M * bonbons I could, but am not :) 1137873388 M * Hollow ok, libvserver-1.0.2 in portage cvs 1137873497 T * Hollow http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.1.0.5, 2.0.1.2 | util-vserver-0.30.209 | libvserver-1.0.2 & vserver-utils-1.0.2 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1137873625 J * Lars ~vaq@you.are.the.lamer.la 1137873670 M * Lars where can i find a .diff file for kernel-2-6-15 ? 1137873705 M * Hollow http://vserver.13thfloor.at/Experimental/ 1137873729 M * Lars http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.1.0.4.diff 1137873730 M * Lars right? 1137873758 J * liquid3649_ ~Viper0482@p5497774A.dip.t-dialin.net 1137873781 M * Hollow yep 1137873783 M * Hollow but that devel 1137873796 M * Hollow if you want stable use 2.0.1.2 1137873817 M * Lars http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.0.1.1.diff 1137873822 M * Lars is the only 2.0 i can find 1137873836 M * Lars oh no sorru 1137873841 M * Lars http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.0.1.2.diff 1137873845 M * Hollow yep 1137873959 M * FaUl is there a delta between 2.1.0.4 and 2.1.0? 1137873971 M * Lars Hmmm can anybody remember in which package make-kpkg is in? 1137873993 M * Hollow FaUl: guess you have to look at http://vserver.13thfloor.at/Devel/ for deltas 1137873999 M * lonewolff kernel-package 1137874002 M * Lars ye 1137874018 M * Lars exec debian/rules DEBIAN_REVISION=custom01 APPEND_TO_VERSION=+vserver KPKG_SELECTED_PATCHES="vserver" INITRD=YES ROOT_CMD=fakeroot binary-arch 1137874018 M * Lars debian/ruleset/misc/patches.mk:104: *** Could not find patch for vserver. Stop. 1137874023 M * FaUl mh, maybe i should update to 2.1.0.5? 1137874052 Q * Lars Quit: 1137874078 M * Hollow FaUl: only the wind will tell you ;) 1137874101 M * FaUl mh, we should realy get bertl into writing changelogs ;-) 1137874110 A * Hollow nods 1137874111 M * bonbons Hollow: Ok, up with 2.6.15-vs2.1.0.4, "vserver start guest" now produces "Failed to restore cwd: Bad address", followed by complain about mounting filesystems 1137874151 M * Hollow hm, is this still with omit-fp? 1137874180 M * bonbons yes, but with dietlibc 1137874194 Q * Viper0482 Ping timeout: 480 seconds 1137874205 M * Hollow hm, could you try without nevertheless? 1137874221 M * Hollow maybe i should filter them for vserver-utils too 1137874323 M * bonbons vserver-utils should link against dietlibc anyhow if libvserver does... so vserver-utils should check how libvserver is compiled 1137874428 M * Hollow with libvserver-1.0.2 and vserver-utils trunk the following should possible: libvserver(glibc) + vserver-utils(glibc), libvserver(glibc) + vserver-utils(dietlibc), libvserver(dietlibc) + vserver-utils(dietlibc) 1137874601 M * _Roey Is there any work on getting OpenVPN to run in a vserver? 1137874645 M * bonbons ok. Is it normal that vserver procfs hide returns 1 and unhide returns 2? (now without omitfp) 1137874651 J * Lars ~vaq@you.are.the.lamer.la 1137874652 M * Lars fs/built-in.o: I funktionen "version_read_proc": 1137874652 M * Lars proc_misc.c:(.text+0x31a5c): undefined reference to `vx_linux_banner' 1137874652 M * Lars make[1]: *** [.tmp_vmlinux1] Fejl 1 1137874653 M * Lars make[1]: Forlader katalog '/usr/src/linux-source-2.6.15' 1137874654 M * Lars Hmmm? 1137874676 M * waldi fix the failed patch 1137874696 M * waldi or use http://137.250.31.225/svn/debian-kernel/local/vserver/linux-2.6 1137874710 M * Lars when i patch it i dont get any errors 1137874739 M * bonbons Hollow: now I get different failure when trying to startup the guest: failed to create network context: File Exists 1137874758 M * Lars Hunk #1 FAILED at 31. 1137874758 M * Lars 1 out of 1 hunk FAILED -- saving rejects to file init/version.c.rej 1137874759 M * Lars waldi 1137874764 M * Lars did: Server9:/usr/src/linux-source-2.6.15# cat patch-2.6.15-vs2.0.1.2.diff | patch -p1 1137874769 M * Lars how come? 1137874810 M * Hollow bonbons: vnflags -S -n -f ~PERSISTANT 1137874813 M * Hollow and try again 1137874870 M * Lars Hollow any ideas? 1137874886 M * Lars will try 2.0.0.1 1137874887 M * bonbons Then it's again the bad address thing 1137874919 M * Lars Hunk #1 FAILED at 31. 1137874919 M * Lars 1 out of 1 hunk FAILED -- saving rejects to file init/version.c.rej 1137874920 M * Lars AGAIN 1137874926 M * Lars using patch patch-2.6.15-vs2.0.1.1.diff 1137874933 M * Lars kernel linux-source-2.6.15 1137874941 M * Hollow maybe the patch is broken.. try to fix the failed hunk 1137874959 M * waldi no, the patch simply not works directly on top of the debian sources 1137874959 M * Lars hehe and how would i do that. 1137874971 M * Lars waldi how can this be fixed then? 1137874993 M * waldi Lars: a person which wants to build patched kernels needs to know that 1137875021 M * bonbons Lars: use vanialla kernel from kernel.org 1137875036 M * bonbons *vanilla 1137875038 M * Lars oh yea dooh sorry guys 1137875536 M * bonbons Hollow: I still get different failure when trying to startup the guest: failed to create network context: File Exists 1137875584 M * Hollow bonbons: hm, could you try to reproduce the bad adress error with UVDEBUG=1 vserver start ? 1137875606 M * Hollow and paste the output somehwere 1137875621 M * bonbons If I can get rid of the network context this should be easy (if possible without reboot :)) 1137875653 M * Hollow vnflags -S -n -f ~PERSISTANT 1137875798 M * bonbons Ok, I have it, now need to restart to get the output to a file :) 1137875886 M * Hollow vkill -x && UVDEBUG=1 vserver start &> bash.debug 1137875887 M * Hollow ;) 1137875896 J * Alissa ali@151.81.2.107 1137875903 M * Alissa hi all =) 1137875904 M * _Roey ello alissa 1137875916 M * _Roey ok, female nickname, you're sure to get responses for whatever you need 1137875925 M * Hollow bonbons: there will be a "vserver release " command in 1.0.3 which does the persistant thing for you 1137875941 M * Alissa _Roey ello is deprecated =P 1137875947 M * _Roey Alissa: :) 1137875949 M * _Roey Alissa: EHLO 1137875953 M * _Roey how about that one :) 1137875963 M * Alissa _Roey ok, it's correctly =P 1137875969 A * Hollow prefers olla! 1137875970 M * Hollow :) 1137875971 M * _Roey Alisser, so you use vserv? 1137875980 M * bonbons ok, I now have the output, will upload it to the usual location: http://homepage.internet.lu/brunop/vsjava.start 1137875987 M * Alissa _Roey yep, on my server =) 1137875999 M * _Roey Alissa: which services does it run? 1137876016 M * _Roey Alissa: we're going to use it for web, dns, mail and maybe othe rstuff. 1137876019 M * Alissa entoo-vserver on emerge.. 1137876022 M * _Roey ahhhh 1137876027 M * Alissa _Roey yep 1137876028 M * bonbons uploaded 1137876031 M * Hollow _Roey the macho 1137876033 M * Hollow :) 1137876035 M * _Roey Hollow: oy nu 1137876038 M * Alissa lol 1137876077 M * Hollow bonbons: well, somehow the bash debugging is output is not in the file 1137876081 M * Alissa _Roey : Linux matrioska 2.6.12-vs2.0-gentoo-r1-matrioska #1 Mon Dec 19 14:37:36 CET 2005 i686 Intel(R) Celeron(R) CPU 2.60GHz GenuineIntel GNU/Linux 1137876085 M * Alissa =P 1137876085 M * Hollow did it appear on your screen? 1137876088 M * _Roey Alissa, Hollow: Do you guys know about the status of OpenVPN on vserver? 1137876092 M * _Roey Alissa: nice! 1137876101 M * _Roey Alissa: hey wait.. matrioska.. you know russian by any chance? 1137876103 M * Hollow gentoo, how lovely ;) 1137876108 M * bonbons yes, bad tee, just taking stdout when it goes the stderr 1137876138 M * bonbons and I never got out how to pipe stderr to another process :( 1137876165 M * Hollow UVDEBUG=1 vserver start foo 2>&1 | tee 1137876222 M * Alissa _Roey I have used matrioska as hostname what the matrioske to their inside has many things... the vserver in this case =) 1137876233 M * _Roey ahhh 1137876252 M * _Roey Alissa: I was going to use Xen until I found that Xen imposes all these requirements (like GRUB, say) 1137876265 M * Alissa _Roey XEN is deprecated... 1137876270 M * Hollow haha 1137876278 M * _Roey Alissa: and besides that, Xen doesnt' offer security if the guest 'computer' is compromised 1137876283 M * bonbons ok, I just took stderr this time 1137876283 A * Hollow points at Xen 1137876285 M * _Roey it becomes like anohter rogue computer on the 'net 1137876290 M * _Roey and NO ONE seems to realize that!! 1137876330 M * Hollow bonbons: did you upload it? 1137876336 M * _Roey VServer, on the other hand, acts like a chroot/jail, where even /root/ on the guest cannot directly access the hardware or even change its IP 1137876342 M * bonbons yup 1137876350 M * Hollow still the same 1137876372 M * bonbons that's what I just saw, trying again to upload 1137876427 M * bonbons This time it's better, and also has stdout 1137876504 M * Alissa uhmmm 1137876534 M * Alissa but does not exist one patch for the vserver, in order to support IPv6? 1137876564 M * Hollow bonbons: ok, please do: cd /opt/vs_java && strace -ff -o trace.log /usr/sbin/vnamespace -E -x 11 -- /usr/sbin/vmount -M -f /etc/vservers/vs-java/init/fstab -m etc/mtab 1137876637 M * Hollow Alissa: no, not yet 1137876652 M * Hollow but there is some work going on to support it 1137876725 M * Alissa Hollow nothing a patch experiments they? 1137876726 M * Alissa =/ 1137876769 M * bonbons ok, uploaded at: http://homepage.internet.lu/brunop/trace.log 1137876783 M * Hollow Alissa: http://www.13thfloor.at/NGNET/ebiederm/2.6.14-rc2/broken-out/ 1137876881 M * Hollow hm.. i'll write a patch for that, just a sec 1137876895 M * bonbons Hollow: getcwd(0,0) = -1 ERANGE, chdir(NULL) = EFAULT (Bad address) 1137876975 M * Hollow yep 1137876988 M * Hollow getcwd(0,0) is lazy 1137877028 M * Alissa Hollow cool! thank's =* =) 1137877108 M * Hollow :P 1137877120 A * Hollow the link list 1137877161 M * Hollow bonbons: http://home.xnull.de/misc/vu_getcwd.patch 1137877188 M * Hollow hm, wait 1137877192 M * Hollow patch is wrong 1137877222 M * Hollow ok, fixed 1137877629 M * bonbons Somehow it does not know PATH_MAX... missing include? 1137877641 M * Hollow ah, yep 1137877670 M * Hollow add #include to both vnamespace.c and vmount.c 1137877777 M * michal` qemu++ 1137877781 M * michal` ^^ 1137877812 M * Alissa qemu is deprecated... 1137877821 M * michal` how so ? 1137877856 M * bonbons retried the vnamespace command, which now fails with 'Bad file number', open(NULL, ...) 1137877895 M * Hollow can you upload the trace again? 1137877907 M * bonbons Hollow: done 1137878025 M * Hollow bonbons: hm, output "vnamespace -E -x 11 -- cat /proc/mounts"? 1137878051 M * Hollow bonbons: if you don't want to upload every single fuzz, we can talk via icq or jabber too if you want 1137878284 M * bonbons Hollow: it lists the mountpoints that are expected to be generated for the guest 1137878297 M * bonbons some of them already multiple times! 1137878305 M * Hollow yeah, i thought so 1137878333 M * Hollow vkill -x 123 1137878335 M * Hollow -x 11 1137878337 M * Hollow of course ;) 1137878341 M * Hollow i always use 123 ;) 1137878367 M * Hollow and (just to be sure): vflags -S -x 11 -f ~PERSISTANT 1137878385 M * Hollow and: vnflags -S -x 11 -f ~PERSISTANT 1137878391 M * Hollow s/-x/-n/ 1137878432 M * Hollow now, we'll try things manually, without the vserver command 1137878454 M * Hollow cd /opt/vs_java 1137878463 M * Hollow vncontext -C -n 11 -f PERSISTANT 1137878472 M * Hollow vcontext -C -x 11 -f PERSISTANT 1137878482 M * Hollow vnamespace -N -- vnamespace -S -x 11 1137878519 M * Hollow strace -ff -o trace.log /usr/sbin/vnamespace -E -x 11 -- /usr/sbin/vmount -M -f /etc/vservers/vs-java/init/fstab -m etc/mtab 1137878536 M * bonbons I managed to break the host's list in /proc/mounts (filesystems are still mounted though) => bug for Bertl 1137878539 M * Hollow and upload the trace again 1137878547 J * KarlNyen ~Karl@S01060050bab2a42d.ok.shawcable.net 1137878550 M * Hollow hm 1137878558 M * Hollow break means? 1137878581 P * KarlNyen Leaving.. 1137878642 M * bonbons No it only lists /, all the others don't appear anymore (had an oops while doing vnamespace -C -x 11 ...) 1137878671 M * Hollow ok, never use this command again! :) 1137878686 M * Hollow you won't need it, really 1137878712 Q * Alissa Quit: 1137878772 M * Hollow bonbons: you will have to reboot i guess 1137878772 M * bonbons well yep :), but will reboot the box to prevent interferences with the oops (the vnamespace -N -- vnamespace -S -x 11 fails with op not permitted) 1137878810 M * Hollow ok, just try the commands above when the machine is back up 1137879034 M * bonbons Hollow: uploaded (looks better this time) 1137879065 M * bonbons seems to be the mtab which was null before 1137879095 M * bonbons (which I may have forgotten on commandline) 1137879141 J * pusling_ pusling@195.215.29.124 1137879142 M * Hollow yeah, seems like it worked 1137879181 M * Hollow ok, now: 1137879194 M * Hollow vnflags -S -n 11 -f ~PERSISTANT 1137879201 M * Hollow vflags -S -x 11 -f ~PERSISTANT 1137879204 M * Hollow vkill -x 11 1137879215 M * Hollow and then try "vserver start vs-java" again 1137879222 Q * pusling Read error: Connection reset by peer 1137879299 M * Hollow i also added a check to vnamespace -C 1137879300 M * Hollow ;) 1137879315 M * bonbons vserver start vs-java was silent this time... 1137879338 M * Hollow yay! 1137879340 M * Hollow it worked 1137879357 M * Hollow vserver enter vs-java 1137879377 M * bonbons but /proc/virtual/11/status reports 1 task though I'm expecting 2: initng and a getty 1137879391 M * Hollow go looking inside what's going on 1137879403 M * bonbons failed to migrate: no such process 1137879412 M * Hollow mhm 1137879419 M * bonbons migrate to network context 1137879454 M * Hollow it fails with network context or normal context? 1137879479 M * bonbons with network context 1137879491 M * Hollow hm 1137879505 M * Hollow is there a network context? /proc/virtnet/11/? 1137879518 M * bonbons note: vserver start vs-java exits with 1 1137879538 M * Hollow hm 1137879549 M * bonbons not only one, two of them! 1137879575 M * Hollow yeah, known bug 1137879648 M * Hollow can you strace the whole "vserver start vs-java"? 1137879654 M * Hollow with -ff -o 1137879669 M * Hollow probably it fails to execv the initng 1137879683 M * Hollow i assume you have added the initng style to vps.sh? 1137879778 M * bonbons yes, that's what I did, but looks like one of the rebuilds has overwritten it 1137879817 M * Hollow bonbons: btw, do you use svn trunk? 1137879840 M * Hollow you can use it anonymously 1137879969 M * bonbons I'm currently using v 1.0.2, but updating to svn might be a good idea (to not reenter fixed bugs) 1137880070 A * Hollow nods 1137880130 M * bonbons the strace last indefinitly... strace -ff -o trace.log vserver start vs-java 1137880157 M * Hollow does the log grow or just stop at some vserver syscall? 1137880167 M * bonbons on: waitpid(-1, 1137880193 M * bonbons more is not yet available in trace.log 1137880311 M * Hollow just cancle the trace for now 1137880331 M * Hollow do you have some guest with sysvinit around? 1137880344 M * Hollow so we can test if that works at least 1137880346 Q * liquid3649_ Remote host closed the connection 1137880370 M * Hollow stracing within contexts is not always a good idea ;) 1137880437 M * bonbons you, just have to switch it's initstyle 1137880594 M * Hollow bonbons: ok, just commited a fix to trunk which should allow to strace vnamespace better 1137880713 M * bonbons updated from svn, building for installing... 1137880739 M * bonbons Make.svn fails with missing "ltmain.sh" ... 1137880780 M * bonbons configure.ac:59: reauired file './ltmain.sh' not found 1137880808 M * Hollow svn up 1137880832 M * Hollow did forget to add libtoolize back 1137880960 M * bonbons how does it decide on dietlibc? 1137881119 M * Hollow look at m4/libs.m4 1137881457 M * bonbons same output at the end, also unterminated line "waitpid(-1, 1137881592 M * Hollow hm, can you upload the trace? 1137881624 M * bonbons you didn't add initng yet as I see in vps.sh (but that shouldn't matter) 1137881689 M * Hollow is teh trace with initng or sysvinit? 1137881713 M * bonbons it's uploaded 1137881873 M * bonbons initng support is not yet in vps.sh, but I did change style to plain 1137881908 M * Hollow hm, very strange.. it seems to hang somehwere in bash 1137881930 M * Lars Ain't it possibel for the virtual server to have a ip that is in another subnet than the main server is in? 1137881964 M * bonbons Hollow: forking often at the end, and one fork does not end... 1137882006 M * bonbons Lars: you can as long as both subnets do not overlap, if they do overlap it may/will cause trouble 1137882068 M * Lars bonbons: The problem is that it's not using the same gateway 1137882085 M * Lars bonbons: is it possibel to define in 0/ip that it should use i a specifyed gatewau 1137882102 J * Thorsten ~Thorsten@dslb-084-058-179-163.pools.arcor-ip.net 1137882139 M * Hollow bonbons: yeah.. somewhere in vps.context setup 1137882151 M * bonbons Lars: different default gateway depending on local IP? 1137882162 M * Hollow i often get neverending straces with contexts.. 1137882169 M * Hollow as i said, it's not the best idea to do 1137882191 M * Lars bonbons: Both ip ranges doesn't use the same gateway is 0/gateway a valid file? 1137882231 M * bonbons Lars: I don't know... check the flowerpage 1137882269 M * Lars hm? 1137882323 M * Lars ?flowerpage 1137882371 M * bonbons Lars: look at the wiki on linux-vserver.org :) flowerpage is the page with doc on util-vserver's configuration (http://www.nongnu.org/util-vserver/doc/conf/configuration.html) 1137882392 M * Hollow .oO(serious documentation ;) 1137882494 M * mnemoc the weed page is serious 1137882502 M * Hollow yeah 1137882513 M * Hollow not that i'm against weed.... 1137882517 M * Hollow ;) 1137882532 M * mnemoc ) 1137882538 M * bonbons Hollow: hmm, there are many tracelogs, will restart an cleanout old ones as there is one per PID... 1137882562 M * Hollow yeah, probably leave out the -ff for now 1137882635 J * mkhl ~mkhl@200-148-41-45.dsl.telesp.net.br 1137882699 M * Hollow bonbons: btw, with the svn version you can now do "vserver release vs-java" instead of the vflags .. persistant thing 1137882723 A * lonewolff heads to be 1137882724 M * lonewolff night all 1137882730 M * Hollow night lonewolff 1137882742 M * bonbons Hollow: yeah, I've just quickly read the strace man-page looking for option to get a little more of the waitpid call, but not found 1137882796 M * Hollow bonbons: you could try again with UVDEBUG=1 and see where it exits with 1 so we can track down this at least a bit 1137882838 M * Hollow btw, i updated the changelog ;) 1137882925 Q * monrad Read error: Connection reset by peer 1137882926 M * bonbons btw, the guest got started, possibly the strace was waiting on the spawned init process! 1137882950 M * Hollow mv head table 1137882997 M * bonbons yeah, that's probably the cause as without strace execution ends immediately (with $? = 1) 1137883018 M * Hollow but it should exit with 0 normally 1137883030 M * Hollow which baselayout-vserver version does the guest have? 1137883143 M * bonbons baselayout-vserver-1.11.13-r1 1137883173 M * Hollow hm, iirc it should work, but i'd update to 1.11.14 to be sure 1137883181 M * Hollow it's ~x86 1137883201 J * monrad ~mikkel@213083190131.sonofon.dk 1137883245 M * bonbons UVDEBUG output is at http://homepage.internet.lu/brunop/vsjava.start 1137883326 M * Hollow hm, seems fine 1137883336 M * Hollow maybe it always returned 1 and i never noticed :P 1137883420 M * bonbons might be :) possibly with similar behavior for procfs, hide and unhide return 1 and 2 don't remember which one returns which value 1137883459 M * bonbons unhide returns 2 1137883496 M * Hollow ok, checking now 1137883733 M * bonbons why does baselayout-vserver-...14 depend on >=sysvinit-2.86-r3 (also ~x86) dependency on one of the patches? 1137883744 M * Hollow bonbons: well, i guess it's as easy as exit 0 in the end 1137883760 M * Hollow bonbons: yeah, the off-by-one patch is for openvz 1137883771 M * bonbons ok 1137883883 M * Hollow i guess the return 1 comes from fs.unlock, but that has to be rewritten anyway, so i tend not to care for now 1137883918 M * bonbons well, let's see :) 1137883936 M * Hollow the bash side of things is more or less a proof-of-concept for now anyway 1137883967 M * Hollow i'm not happy with most if its code ;) 1137884047 M * bonbons ther's lots of things that can be improved/extended, an some manpages for the c code would be nice, will eventually try to write something 1137884076 M * Hollow bonbons: i already started man pages, they're not distributed yet, but in svn trunk/man 1137884124 M * bonbons ok, so will look there 1137884165 M * Hollow btw, thanks for your time and debugging on vserver-utils :) 1137884210 M * bonbons That's no problem, thanks for your patience :) 1137884254 M * Hollow bonbons: if want to contribute in long term i can also give write access to you 1137884319 M * Hollow guess will leave to bed in a few minutes *yawns* 1137884371 M * bonbons we will see how it develops, hard to say how much time I will have, but vserver is quite useful, until then I will help out via irc/uploads 1137884398 Q * Doener Quit: Leaving 1137884412 M * bonbons for me the same thing applies midnight 1137884424 M * Hollow yup, will be busy with school too in the next few months, so every minute spent on vserver-utils is a well-spent one ;) 1137884489 M * bonbons for me it's not more scholl, but work, sports and worse, public transports 1137884511 Q * pusling_ Read error: Connection reset by peer 1137884518 P * meandtheshell 1137884528 J * pusling pusling@195.215.29.124 1137884549 Q * entroposcope Ping timeout: 480 seconds 1137884594 J * denote toor@sandbox.iax.be 1137884669 M * denote can someone point me to some docs about ipv6 + vserver ? 1137884787 M * Hollow denote: vserver does not work with ipv6 atm, but there is some early work going on 1137884812 M * bonbons denote: nothing there yet, ngnet is proof of concept, an legacy network API is IPv4 only at the moment 1137884833 J * Smutje_ ~Smutje@xdsl-87-78-0-172.netcologne.de 1137884843 M * Hollow bonbons: with vserver-utils you do not need legacy net api anymore btw 1137884862 M * Hollow it fully supports the new api 1137884881 M * bonbons ok, bad term :) I meant the chbind technique 1137884892 M * Hollow yup, that's the legacy api ;) 1137884898 M * Hollow alrighty... off now 1137884904 M * Hollow night all 1137884913 M * bonbons night Hollow 1137884924 J * titoo ~titoo@dh3cp26.kt.dtu.dk 1137884928 M * titoo hello 1137884939 Q * Smutje Ping timeout: 480 seconds 1137885244 T * * http://linux-vserver.org/ | latest stable 2.01, 1.2.10, 1.2.11-rc1, devel 2.1.0, exp 2.1.0.5, 2.0.1.2 | util-vserver-0.30.209 | libvserver-1.0.2 & vserver-utils-1.0.2 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the wiki, and we'll forget about the minute ;) 1137885244 T * Hollow - 1137885246 M * denote nvm, make died on my slack box 1137885250 Q * pusling Ping timeout: 480 seconds 1137885261 M * titoo I try to setup another vserver on the new host to test 1137885262 M * bonbons denote: yes they do conflict in many ways 1137885262 Q * Bertl_oO Remote host closed the connection 1137885272 M * bonbons denote: includes, binaries 1137885281 M * titoo gdm: how can I check that? 1137885291 J * Bertl_oO herbert@212.16.62.52 1137885331 M * bonbons "ip addr list" (iproute2) from the host, an check that the ip is on the right interface 1137885365 M * gdm titoo: look at /etc/vservers/*/interfaces//ip 1137885389 M * gdm is the network card, ie eth0 eth1 etc 1137885394 M * titoo is the here important ? 1137885400 M * titoo ok 1137885406 M * titoo so if eth1 1137885415 M * titoo need 1 as the number? 1137885423 M * gdm but also look at /etc/vservers/*/interfaces//dev 1137885430 M * gdm and check it is what you expect 1137885443 M * gdm titoo: yes, 0 or 1 or something like that 1137885450 J * pusling pusling@195.215.29.124 1137885468 M * titoo ok I check 1137885536 M * gdm root@HOST:~# cat /etc/vservers/VSERVERNAME/interfaces/0/ip 1137885536 M * gdm ip.ad.re.ss 1137885536 M * gdm root@HOST:~# 1137885548 M * titoo here it is ok 1137885570 M * gdm root@HOST:~# cat /etc/vservers/VSERVERNAME/interfaces/0/dev 1137885571 M * gdm eth0 1137885571 Q * denote Quit: leaving 1137885602 M * titoo on the host I have eth1 working 1137885617 M * gdm well, then it should be interfaces/1 1137885633 M * gdm instead of interfaces/0 1137885635 M * titoo ok 1137885638 M * gdm also, did u see this: 1137885638 M * gdm < bonbons> "ip addr list" (iproute2) from the host, an check that the ip is on the right interface 1137885650 M * titoo yes checked 1137885683 M * gdm probably need to stop, change, then start vserver too (if it's running) 1137885725 M * titoo the output: http://pastebin.com/516756 1137886025 M * titoo and now I have only interfaces/1/ 1137886025 M * titoo with eth1 in dev 1137886025 M * titoo and the ip in 1/ip 1137886025 M * titoo ifconfig within the vserver provides no ip address 1137886025 J * ewqeqw ~wwqwe@85.101.245.204 1137886025 M * Thorsten Hi everyone 1137886025 M * titoo I have updated: http://pastebin.com/516762 1137886028 M * titoo with netstat -l & ifconfig & ip addr list from within the vserver 1137886165 M * Thorsten I want to try vserver and the latest stable kernel (2.6.15.1). Does anyone know if there is already a verserver kernel-patch available? 1137886210 M * bonbons the development patch agains 2.6.15 should probably work 1137886270 M * titoo so has anybody an idea on the procedure to follow to debottleneck that? 1137886331 M * bonbons from host, ping hte guest, and check if it works. Also try setting both addresses to the same mask 1137886358 M * Thorsten Mmm, this one? http://vserver.13thfloor.at/Experimental/patch-2.6.15-vs2.1.0.5.diff 1137886372 M * titoo from the host I can ping the guest it works! 1137886386 Q * ewqeqw Quit: | ® | KELEBEK ScripT v3.1 | ® | Mutlaka Download Edin! | © | Www.GencSite.Net | © | irc.gencsite.net | ® | 1137886413 M * titoo bonbons: how can I set the same mask? 1137886415 M * bonbons Throsten: yes 1137886468 M * bonbons change the value in interfaces/x/prefix 1137886470 M * Thorsten Ah, thx. And I can apply it direclty on the kernel sources or do I have to apply other patches at first? 1137886482 M * titoo bonbons: thanks 1137886520 M * bonbons should work directly on vanilla sources 1137886575 M * titoo bonbons: it works 1137886590 M * titoo I can ping from outside 1137886596 M * titoo it looks like 1137886631 Q * pusling Read error: Connection reset by peer 1137886633 J * pusling_ pusling@195.215.29.124 1137886846 M * Thorsten I upgraded this week to kernel 2.6.14.6 with patch-2.6.14.4-vs2.1.0. Now I noticed that my ftp server (which is inside a verser) delivers broken files. Is this a known issue? 1137886943 M * bonbons I've nerver heard of such an issue... are all files broken, how does the ftp-server send the files; sendfile()? 1137887010 Q * nokoya Read error: Connection reset by peer 1137887017 M * titoo bonbons: youpi! It works! http://capec.kt.dtu.dk/ 1137887026 J * nokoya young@hi-230-82.tm.net.org.my 1137887027 M * titoo running from the vserver 1137887030 M * titoo thanks a lot 1137887037 M * Thorsten It seems if the file a file is abcd I get abcdabcdabcd..., 7MB files are somtimes > 100MB 1137887064 M * Thorsten I don't know how prpftpd sends its file. 1137887111 M * Thorsten I will at first try the newest kernel and then try to debug it further if the problem persists, ok? 1137887119 J * rainbow ~nd@host36-106.pool878.interbusiness.it 1137887128 P * rainbow 1137887149 M * bonbons a long time I didn't run it anymore, now I'm using oftpd, maybe there's a config option for sendfile() [sendfile is known to be sometimes broken] 1137887209 Q * Lars Read error: Connection reset by peer 1137887266 M * titoo bonbons: Really thanks a lot I was upgrading an old vserver host (3 years old no problem at all) from 2.4 to 2.6 and started to be a little annoyed 1137887321 M * bonbons migration from 2..4 to 2.6 is not that easy, need to take care about utilities, get kernel config right :) 1137887357 M * titoo and relearn vserver, the config has completely changed since last time :) 1137887379 M * titoo 1 am, you saved my night :) 1137887384 M * titoo good night 1137887399 P * titoo Thanks again bonbons 1137887433 M * bonbons yes, an from util-vserver to vserver-utils the same step (but that's something to start in the next months) 1137887972 Q * bonbons Quit: good night